The full text is 4023 words in total and takes 13 minutes to read
Where does it exist operate , where needed Data analysis , WeChat Public account The same goes for operations.
Although the public account has been criticized since the beginning of 2017, including that the average image and text opening rate has fallen to only 4% to 5%, it still cannot prevent enterprises and we media people from entering the market. Why is this?
In short, because the products that can really replace the public brand have not yet appeared.
1、 Is the public account worth doing?
Specifically, the public account has three advantages that are hard to shake:
First, The subscription system of public accounts. It establishes a relatively strong relationship between content producers and fans, and can guarantee an information arrival rate of more than 97%. Less than 3% of the remaining fans can't reach it because they haven't logged in for a long time, have been blocked, or have set up not to accept push messages. Compared with other We Media platforms such as Toutiao and Jianshu, although they also have subscription functions, due to the limitations of user base and push mechanism, they need to rely more on recommendations to obtain traffic.
Second, WeChat traffic advantage. There is no need to explain this too much. WeChat, which has broken the 1 billion mark of monthly activity, has its own advantages in traffic.
Third, Easy to develop. WeChat is inherently suitable for helping start-ups to verify the feasibility of their products. Developers only need to master website development technologies such as Html5, Java and PHP to create various webAPPs (web applications) based on WeChat ecology.
I also have a deep understanding.
In the process of operating a tool type product, due to the lack of investment in public account operation in the early stage, three problems were caused: first, it was very passive in BD and could not exchange good resources; Second, users can't settle down, which means that there is no channel for information distribution and user interaction; Third, the original content was plagiarized, and the articles originally posted on the official website were copied, pasted or washed by competitors to their public accounts and declared original.
Based on the above analysis, public accounts are very important to products, operations and users. In many modules of public account operation, Data analysis It is also the most important, the means to find problems, and the basis for operational strategy adjustment.
Next, I will share my ideas and methods of doing public account data analysis.
First of all, the public account has been born for many years, but there is still no unified rule or algorithm for the definition of its data analysis. Different third-party platforms may differ from company to company, because different companies have different definitions and focuses on data, just like the definition of active, some products are active once users log in, However, some products stipulate that users must have key business behaviors to be active.
2、 How to do data analysis of public accounts?
Analysis tool: Excel 2013 version or above
In fact, more than 90% of the data analysis of operational work can be completed efficiently by Excel. But the premise of efficient completion is that you need to master the skills of using Excel, be proficient in operating common functions such as pivot, chart, vlookup, if, sumifs, and countifs, and also have the idea of data analysis based on Excel, be familiar with the meaning of each data indicator, and analyze with the destination.
1. Establish the concept of source data table, sub table and summary table
Ps: Old operators can skip this part.
First popularize three concepts: source data table, sub table and summary table.
Source data table refers to a table form with column header, one cell corresponds to one data, data is not calculated, and no merged cells exist. The reason for this is to follow the "rules of the game" of Excel in data processing, and use the rules of the game to improve analysis efficiency.
For example, in the source data table, one cell corresponds to one data, and there are no merged cells; One concept, one column of data, no mixed concept; There is no blank line.
When designing the source data label, it should be single, standard and concise. Singularity means that a concept should have an exclusive column. Do not try to merge it, or it cannot be split during analysis. The format of each column of data is required to be uniform, otherwise the analysis results will make mistakes in summation and counting items. Conciseness means that there should be no redundant fields, which is consistent with the principle of PPT design - if unnecessary, no entities should be added.
The source data table is the source of all data analysis. Sub tables and summary tables are derived from the source data table. When the source data table changes, the data of sub tables and summary tables will also change. Therefore, I usually regard the source data table as a separate table. If the data is not updated, the source data table will not be changed.
After the source data table is clarified, the sub table and summary table become simple. A sub table, as its name implies, is a table that is analyzed and extracted from the source data table. It is used to analyze one or more items of data. The most common Excel function used to make a sub table is PivotTable. A summary table is a summary of various data in a sub table.
With the source data table, sub table and summary table, a complete set of data analysis can be made. This method is not only applicable to the data analysis of public accounts, but also to the data analysis of activities, users, websites and APP.
2. Source data and analysis methods to be collected
In the WeChat official account, the data most worthy of deduction and analysis are mainly in three dimensions: graphic data, user data and menu data. Among them, graphic data and user data can most intuitively reflect the operation results of a company account. Naturally, these two dimensions of data have also become the most concerned data, so many new media people will be proud of 10W+and a large number of fans.
However, they are just the tip of the iceberg. There are still many important data in the public account that outsiders cannot see and pay little attention to, such as user source channel data, menu click data, image and text reading data distributed by source, reading volume distribution data by time, news statistics, secondary communication data, and so on.
1) Graphic data
The source data to be collected include: reading amount, like amount, comment amount, sharing amount, collection amount, number of users delivered.
According to the six source data, we can get some reference indicators:
(1) Reading rate , also known as the open rate, it is equal to the reading volume divided by the number of users delivered, and then multiplied by 100%. The role of the reading rate is to measure the reading quality of different public accounts in the same period or the same public account in different periods fairly. Because we will find that when we compare the reading quality of different public account operations, because of different fan equivalents, we cannot judge by the single indicator of reading volume; Or compare the reading quality of the same public account in different periods. Because its number of fans in different periods is different, it can not be judged by the amount of reading.
(2) Like rate , which is equal to the number of likes divided by the amount of reading, and then multiplied by 100%. The likes rate reflects the user's acceptance of the content.
(3) Sharing rate , which is equal to the amount of sharing divided by the amount of reading, and then multiplied by 100%. The sharing rate also reflects the degree of user recognition of the content, and is also the driving force of the secondary dissemination of content.
(4) Collection rate , which is equal to the number of collections divided by the amount of reading, and then multiplied by 100%. The collection rate reflects the user's recognition of the content. But the high collection rate may also be because the article is too long for users to finish reading.
(5) Comment rate , equal to the number of comments divided by the number of readings, and then multiplied by 100%. The comment rate can reflect the stickiness and activity of a public account user.
Source data table reference of graphic data
The following source data representation style is applicable to all types of public accounts.
Follow the specifications of the source data table, do not have merged cells, and do not habitually write the table in the first row (it is recommended to write it in the navigation tab at the bottom).
Click to view the large picture, and reply to the keyword "table" in the public account dialog box to obtain the Excel table
With this source data table and the pivot function, you can not only pull out the analysis charts of various historical data at any time, but also analyze the user's content preferences very quickly.
In addition to the image and text data mentioned above, there are four data that need to be understood but do not need to be extracted for analysis, namely:
(1) Readings by source
The reading sources of most public accounts are mainly dialog boxes, friend circles and friend forwarding. The proportion of them can reflect the situation of this account.
Distribution of reading volume by source
For example, the proportion of dialogs is high, probably because the quality of the articles is very poor, the users walk away at a glance, and no one is willing to share, or the articles do not have the dissemination attribute, and there is not enough motivation to drive users to share. For example, if the proportion of other channels suddenly increases, it is likely to be the amount of brush.
(2) Data of reading distribution by date
The amount of reading distributed by date is very interesting. It is usually an L-shaped curve. The amount of reading in the first two to three days after the release usually accounts for more than 70% of the total amount of reading.
Distribution of readings by date
The L-shaped curve is actually the "long tail curve" that plagues most small and medium-sized brands in the Internet era.
(3) Data on the distribution of reading volume by gender
Gender is one of the most important factors in the operation of public accounts, because men and women differ greatly in aesthetics, reading habits and preferences for content.
Distribution of reading volume by gender
The data here, combined with the data on the proportion of men and women in the user attributes, allows you to choose the layout style of your public account - more feminine, more masculine, or more stable and neutral. You can also adjust the topic selection from this.
(4) Readings by region
Through the regional distribution data, you can know which provinces and cities your fans are mainly distributed in.
In social relations, region is one of the important reasons that drive people to pay attention, share and talk, because it is human nature to pay attention to their birthplace and hometown. Many public accounts make use of this to write articles in various cities or provinces and quickly circle fans.
2) User data
The source data to be collected include: number of new followers, number of cancelled followers, net increase in number of followers, cumulative number of followers, and user source channel data.
User data is relatively simple, which is mainly divided into two parts: first, user growth data, which changes every day, needs your attention; Second, user attribute data is equivalent to an incomplete user profile. Although it actually changes every day, it does not need to be paid attention to every day because the range of change is small.
Therefore, the only data we need to collect is user growth data. User growth data also includes daily user increment, decrement, cumulative volume and source channels.
New followers, cancelled followers, net increase followers and cumulative followers can be directly exported in the background of the public account. Focus on trend and important event nodes during analysis.
The user source channel data can help us analyze the user source proportion of each channel. Combined with the historical data, we can see whether the data of a channel is on the high side or on the low side, and also let us see which channels need key investment and which channels can be abandoned.
For the channel that needs to be focused on, continue to deepen, focus on users' attention to the path on this channel, find ways to shorten and simplify this path, and then expose it as much as possible.
Source Data Table Reference for User Data
The user growth data table can be downloaded directly from the public account background, and can be made into a source data table after simple processing. If the user comes from channel data, it needs to be entered manually.
As it is trivial to extract channel data manually, whether to extract user source channel data mainly depends on the strategic position of the public account in business. If your company all in is on the public account and needs to find any possible growth point, it is necessary to pull it out for analysis. If your company's main business is far away from the public account, you can only use the public account as one of the distribution channels without analysis.
Click to view the large picture, and reply to the keyword "table" in the public account dialog box to obtain the Excel table
PS： If you want to quickly obtain accurate data of a channel, you can use the QR code with parameters.
3) Menu Data
The source data to be collected includes: number of menu clicks and number of menu clicks
The public account menu, like the bottom menu navigation of APP, needs to have a clear information architecture, so that users can find the functions they want without thinking, and spend the least number of clicks. This means that a balance needs to be found between information classification and high-frequency functions.
Information classification requires that you have sufficient knowledge of the company's business, so that it is possible to avoid omission and repetition; On the one hand, the high-frequency function is based on the business focus, and on the other hand, it is based on the menu click data.
What about menu data?
Look at the number of clicks per capita. The more clicks per capita, the higher the frequency of this function, and the menu level needs to be advanced.
Second, look at the trend. Specifically, it refers to the change trend of the total number of menu clicks, which is one of the feedback indicators of user activity. In order to evaluate the health of the trend, you can put it together with the trend line of user changes and compare the growth rate. The growth rate of the total number of menu clicks is greater than the growth rate of users, which means that users are more and more active, and vice versa.
Source Data Table Reference for Menu Data
Suppose an official account has two first level menus, marked A and B respectively. A contains two second level menus, marked a1 and a2, and B also contains two second level menus, marked b1 and b2. Then the style of the source data table is:
Click to view the large picture, and reply to the keyword "table" in the public account dialog box to obtain the Excel table
This set of data analysis method is the result of my exploration and constant correction in practice. It is not guaranteed to be effective for all people and all numbers, because in the whole industry, there are also some problems that different people have different views on, such as the value of reading volume and the uncertainty of other reading sources. But these are just points, which do not affect the area and trend, nor mislead the operation strategy.
Although the method is simple, it can diagnose most of the problems existing in the public account, and after being familiar with it, it can also transform many practical tips, such as reverse calculating the total number of fans of a public account according to its reading amount; For example, estimate the amount of zombie powder; For example, calculate the user's retention rate for the next day, the 7-day retention rate and the 30-day retention rate; For example, judge whether there is brushing behavior according to the data; For example, distinguish between garbage number and valuable number.
If you want to know this, please leave a message and communicate with me.
Head picture by Kang Hee Kim
